# User Portlet Richard Frost
Discussions
Rohit, Thank you for the tips. I have been misunderstanding ':=' as 'define', and also was unaware of RandomSample[] - which I'm reading has the desired properties.
Hello Henrik, thank you for the clarification. This is actually a very elegant way of computing the solution in a simple way, without having to calculate any coefficients! Funnily (and understandibly), it is converging faster when n gets larger. ...
Radicals are tricky. Equations with radicals, doubly so. I have learnt from your nice example. I hope you have too.
Thanks, it worked.
It is not that there are drawbacks to k-means, but rather there are drawbacks to implementations that ignore multiplicity. Part of this is due to the influence of of older k-neighbor algorithms, designed for compiler construction in a single threaded...
Is there a list of logical rectangular (over-determined) matrices where the pseudoinverse can be solved by a simple algorithm rather than the Mathematica Function PseudoInverse (Moore-Penrose method)? I seem to recall a page on Wikipedia that...
Thank you Kuba for your quick reply plus examples. So "& @" now makes sense, a prefixed pure function. NumberForm is a good idea too -- it gets to the heart of what I wanted. Quantity[] is also desirable, but upon export in tabular form does not...
"I would say PT is the geographic region that switches between PST and PDT". The actual definition has PDT as a zone to accommodate portions of PST that do/did/will not change during the year. But I endorse the idea returning a value other than a...
Hello all, What do you think about the values given by "Wolfram Alpha" in the cases: 1) sum_{k=n}^1 k = -9 for n=5 ? 2) sum_{k=5}^1 k =0 ? Thank you very much! Iulian
I won't confess to a "whim" but anything from 31 and on seems to provide a smooth enough prediction (a prediction much smoother than the observed data). Also using an odd number gets one the appropriate associated time (the middle of the time...